diff --git a/mpwobj.py b/mpwobj.py
new file mode 100644
index 0000000..94d0da7
--- /dev/null
+++ b/mpwobj.py
@@ -0,0 +1,137 @@
+import struct
+
+class MPWObject:
+    def __init__(self):
+        self._list = []
+        self._dict = []
+        self._backdict = {}
+        self._dict_idx = 200
+
+    def __bytes__(self):
+        dest = bytearray()
+
+        for chunk in self._list:
+            dest.extend(chunk)
+            if len(dest) & 1: dest.append(0)
+
+        return bytes(dest)
+
+    def _ensurename(self, name):
+        # get the ID of this name from the dict
+        # If nonexistent, then add it
+        # i.e. idempotent
+
+        try:
+            return self._backdict[name]
+        except KeyError:
+            self.putdict([name])
+            return self._backdict[name]
+
+    def _quickappend(self, *bytelist):
+        self._list.append(bytes(bytelist))
+
+    def putfirst(self):
+        self._quickappend(1, 1, 0, 2)
+
+    def putlast(self):
+        self._quickappend(2, 0)
+
+    def putdict(self, items):
+        dest = bytearray()
+
+        dest.extend([4, 0, 99, 99])
+        dest.extend(struct.pack('>H', self._dict_idx))
+
+        flag = False
+
+        for item in items:
+            flag = True
+            dest.append(len(item))
+            dest.extend(item.encode('ascii'))
+
+            self._backdict[item] = self._dict_idx
+
+            self._dict_idx += 1 # ID of the *next* thing
+
+        if not flag: return
+
+        struct.pack_into('>H', dest, 2, len(dest))
+        
+        self._list.append(dest)
+
+    def putmod(self, name='#0001', segname='Main', flags=(1<<7)+(1<<3)):
+        modid = self._ensurename(name)
+        segid = self._ensurename(segname)
+        self._last_mod_id = modid
+
+        self._list.append(struct.pack('>BBHH', 5, flags, modid, segid))
+
+    def putentry(self, offset, name):
+        entid = self._ensurename(name)
+
+        self._list.append(struct.pack('>BBHL', 6, 1<<3, entid, offset))
+
+    def putsize(self, size):
+        self._list.append(struct.pack('>BBL', 7, 0, size))
+
+    def putcontents(self, data): # in multiple chunks please!
+        done = 0
+
+        while done < len(data):
+            this_time = data[done:done+30000]
+
+            header = struct.pack('>BBHL', 8, 1<<3, 8 + len(this_time), done)
+
+            self._list.append(header + this_time)
+
+            done += len(this_time)
+
+    def putcomment(self, cmt):
+        cmt = cmt.replace('\n','\r').encode('mac_roman')
+        if len(cmt) & 1: cmt += b' '
+
+        dest = bytearray()
+        dest.extend([3, 0])
+        dest.extend(struct.pack('>H', len(cmt) + 4))
+        dest.extend(cmt)
+
+        self._list.append(dest)
+
+    def putsimpleref(self, targname, width, *offsets):
+        offsets = list(offsets)
+
+        if width == 2: # of the operand field, in bytes
+            flags = 1 << 4
+        elif width == 4:
+            flags = 0
+
+        flags |= 1<<3 # longwords in the offset list!
+
+        targid = self._ensurename(targname)
+
+        dest = struct.pack('>BBHH', 9, flags, 6 + 4 * len(offsets), targid)
+        dest += b''.join(struct.pack('>L', o) for o in offsets)
+
+        self._list.append(dest)
+
+    def putweirdref(self, targname, width, *offsets):
+        # Assumes that you've already put -offset at offset
+        offsets = list(offsets)
+
+        if width == 1:
+            flags = 2 << 4
+        elif width == 2: # of the operand field, in bytes
+            flags = 1 << 4
+        elif width == 4:
+            flags = 0 << 4
+
+        flags |= 1<<7 # difference calculation
+        # flags |= 1<<3 # longwords in the offset list!
+
+        targid = self._ensurename(targname)
+
+        dest = struct.pack('>BBHHH', 10, flags, 8 + 2 * len(offsets), targid, self._last_mod_id)
+        dest += b''.join(struct.pack('>H', o) for o in offsets)
+
+        self._list.append(dest)
+